Agent
Reviews changed files for a specclaw change across 10 quality dimensions (correctness, security, YAGNI/simplicity, one-liner opportunities, naming, complexity, test quality, design adherence, scope creep, dead code). Produces review-report.md with BLOCK/WARN/NOTE findings and an…
Agent
Party-mode panelist — structural fit. Attacks duplication of existing mechanisms, wrong seams, blast radius at merge time, under-specified contracts, and untestable designs. Read-only; returns findings as its final message. Seated at every tier.
Agent
Party-mode panelist — problem fidelity. Attacks whether the proposal's stated problem is the real problem and whether its evidence is real. Read-only; returns findings as its final message. Seated at tier standard and above.
Agent
Sizes the adversarial review panel for a specclaw proposal. Reads proposal.md and returns a JSON object with a depth tier (thin/standard/deep), domain flags, a one-sentence rationale, and three depth signals. Runs inside /specclaw:propose when party.enabled is true and no tier override is set; specclaw-party turns the…
Agent
Party-mode panelist — value per token. Attacks scope size, cost, the cheapest variant that captures most of the value, the do-nothing option, and the shipped defaults. Read-only; returns findings as its final message. Seated at every tier.
Agent
Party-mode panelist — trust boundaries and misfire behaviour. Attacks hostile input, fail-open failure modes, over-broad permissions, and irreversible runtime effects with no recovery path. Read-only; returns findings as its final message. Seated at tier deep or when a security domain flag is set.
Agent
Party-mode panelist — the long horizon. Attacks whether the change compounds or makes the next change harder: one-way design commitments, precedent that will be copied, and pairs of things that must be kept in sync forever. Read-only; returns findings as its final message. Seated at tier deep.
Agent
Interactively co-authors a high-quality spec.md for a specclaw change. Walks the user through the spec template section by section, applying named brainstorming techniques (5 Whys, Jobs-to-be-Done, Inversion, Pre-mortem, MoSCoW, Concrete-example probe) and challenging vague or untestable requirements before writing…